Advanced Metering Systems Infrastructure:
Accuracy needs must go beyond the meter
Advanced Metering Infrastructure (AMI) continues to be a very active and topical area in smart grid circles. However, the California and Texas media have picked up on customer concerns of the impact that these new devices will have on their bills, and surfaced apprehensions about the “accuracy” of these new solid-state devices. The current results of forensic activity have shown that, in general, meters are very accurate measurement device; however, activities that surround the meter-to-bill process can lead to many possible opportunities for error.
KEMA has been involved with AMI for the past six years, while our roots in certification and testing span decades. Gaining a perspective that includes technologies, people, and processes is critical to form a balanced viewpoint of this issue. In response, KEMA’s expert AMI team, which is actively involved with a number of major projects, recently conducted a “roundtable” discussion to gain insights on the concerns of metering and metering systems’ accuracy.
Members of the team, most of whom are subject matter experts (SMEs) in a specific component of AMI / smart grid, shared pertinent questions and answers taken from major recent projects that were surrounded by customers’ speculation. Following is an excerpt of their discussion about AMI areas that need to be examined, as well as their shared insights into testing and analysis components to use prior to implementing smart meters.
Ron Chebra, Director of AMI: Given the recent awareness and concern about meter accuracy in the press, are there aspects of AMI that need to be examined?
Harry Stephey, AMI Technology Assessment SME: Metering now involves so many more elements than just the meter. Given that AMI also includes an end-to-end system of meter communications devices, networks, collectors, information transport, and back-office processing systems, utilities and their trusted advisors should focus on billing process accuracy, rather than exclusively on meter device accuracy.”
Afshin Tajian, Consultant: This is very true; the entire process is much more complex. AMI systems can provide more than consumption information for billing. The can often gather other valuable information from the field such as phase angle and demand. However, today’s American National Standards Institute (ANSI) standards only look at meter-centric issues, such as accuracy over temperature range. New standards need to explore the holistic metering process.
Craig Bialy, AMI Implementation SME: Today’s meters can capture consumption through the use of multiple registers, rate tables, and meter tables, which creates an opportunity for greater risk insofar as how these variables are assigned and programmed into the meter. It is no longer sufficient or adequate to make sure that a new AMI device has the same meter form and multiplier when installing it. More diligence is necessary to ensure that all of the meter’s elements are properly configured.
Terry McDonald, Principal Consultant: This is especially true when the entire population of meters is changed during deployment. It places a significant burden on the meter database within the utility to ensure that data are current, accurate, and translatable into the solution set that is being deployed. Often there are situations when correcting discrepancies due to incorrect data can have significant impacts on billing. If some of the key billing determinant factors have been incorrect for years, making them current can result in much different bills, independent of installing new meters.
Brian Pugliese, AMI Quality Management System, SME: Because smart metering, i.e. AMI, is gaining a public profile and is also one of the few areas of smart grid that actually touches every customer, there can be a heightened awareness of change. This opens an opportunity for utilities to become more proactive. When you think of the process before automated meter reading (AMR), customers rarely saw the meter reader walking around with a handheld unit.
The process of gathering consumption information to render a bill was an internally focused operational step in the meter-to-bill process. With all the changes taking place in the field and the benefits of a smarter grid building market awareness, it is likely that those customers who have never called their utility about a billing issue, will now show a greater interest in the potential changes this technology will have on their experiences with their utilities and their bottom lines.
Fred Dorow, Principal Consultant: Today’s advanced solid-state meters are growing in maturity. Often the end-to-end meter-to-bill process consists of integrating a number of individual vendor offerings into a holistic system to replace a time-tested manual or semi-automatic process. Both the integration of new technologies and the implementation of new supporting processes leave the potential for errors or inconsistencies.
McDonald: In the past, mistakes have happened—but in relatively low numbers spread over an inventory of meters that varied in age and performance. Every utility has records of incoming high bill complaint calls, many of which resulted from errors that happened in the meter reading process. However, the current risk is that with the mass rollout of meters and the heightened attention the media has given to certain exceptional situations, the wary public will focus on its energy bills with a renewed interest and deeper scrutiny.
Dorow: Likewise, the automation of the reading process now creates a new opportunity to normalize the billing day cycle. No longer will there be the variations in the monthly cycle that exist in the manual process, with billing cycle days that fall on weekends. Short-day cycles set expectations of lower bills; long day cycles can create the perception of increased consumption.
Tajian: Today’s solid-state meters are achieving higher class standards for accuracy. There is little doubt that these devices are capable of achieving new levels of performance. Not only can they measure more accurately, but also they can provide an abundance of additional information that utilities can use to cross-check and validate the data received. Interval data elements can be accumulated to provide totalized readings; logs exist that provide insight into many events, and so forth. What appears to be lacking is the application of all this information in a meaningful way.
Stephey: Recently, I have seen increased efforts by the utilities to educate customers. The level of outreach is substantial. New opportunities exist for utilities to connect with customers and to engage them before, during, and after the meter change process.
Bialy: Getting ahead of the customer wave is very important. Anticipating their concerns is as critical as responding to their needs. Simple facts, such as how AMI reduces estimated reading and limits unaccounted-for consumption, are all helpful and valuable preventative measures in tuning the customer to a new way of thinking about energy, consumption, and costs.Creating call centers to be responsive to customer concerns and complaints is also critical.
McDonald: There is also a need to understand where customers are coming from. Frequently, the most vocal customers that participate in pre-implementation workshops or open forums fall into two outlier classes: those enthusiasts who cannot wait to adopt and leverage new technologies; and those who are vehemently against changes from the status quo. This leaves a wide gap that comprises the mass population of customers—those who have no particular drive to take interest in smart meters, or any other innovation that they cannot control. This target group can be easily influenced by those on the fringe.
BillPeak, Envision Center Manager: This is particularly true if you depend too heavily on the demographics of a given area. You cannot assume much about how people will react to new technologies.
Bialy: This middle gap can represent over 75 percent of the population. It consists of customers who neither know about, nor have interest in smart metering.
Chebra: So given all of these complexities of technology, customer behavior, and so forth, how do you ensure that sufficient tests and analyses are done before going live?
Tajian: Testing has many aspects and levels. Meter accuracy tests are well defined in the ANSI specifications. However, this is only part of the testing that needs to take place. Meter manufacturers need to do extensive integration tests, especially in light of variations that can exist among different suppliers of communication cards that can be installed.Utilities must now include more system-wide tests beyond the meter test bench, including communication systems, message monitoring, traffic performance, and so forth. Frequently, future capabilities of these meters go untested, such as operation of the service disconnect relay under full load, power supply, etc. Smart meters of today are more like personal computers than rotating discs and magnets, and must be treated as such.
Pugliese: Sometimes it appears as if we are trying to operate in a fully mature and stable market. Adopting the attitude that AMI is a commodity purchase, which will be awarded to the lowest bidder, may satisfy a near-term financial benefit, but it may not be the most prudent choice over time. Smart metering will require a relationship beyond the supplier/purchaser, which changes the rules of the game. KEMA has been involved with a number of supplier audits on behalf of our utility clients.
One thing in common with all these exercises is that the level of engagement between the utility and the vendor is significant, enduring, and takes on a partnership characteristic. Process management, source inspections, and compliance with industry standards are relatively easy metrics to track and follow. Being part of a team that is trying to achieve mutual success is a mindset change that must be followed for long-term success of utilities and their customers.
Stephey: This is particularly true as variations of solutions can follow many different approaches. In one case, you might have a relatively dumb meter that interfaces with a smart communications module. In other cases you can have a smart meter that interfaces with a dumb communications module. Although both situations can result in the similar functionality, the level of testing and dependence of the communications modules vary substantially, and they place a level of dependence on a system level test that may be shared by two vendors.
Bialy: This further amplifies the need to do vendor audits prior to the selection process, and most importantly, during the delivery phase of a program. While no one suspects that vendors are deliberately shipping defective or inadequately tested units, it is essential that utilities have an active and continuing role in the delivery process.
Tajian: There also seems to be a lack of industry tracking for failure modes of new smart meters. Experience has shown that traditionally, meters follow the bathtub effect—infant mortality has a high percentage of failures attributable to bad voltage sensors, power supplies, and LCD displays. However, at the other end of the curve, there are few records or statistics that one can find to delineate what is the root cause or failure mode effect. Often these weak links can be stressed through accelerated life testing (ALT) to mitigate concerns.
Stephey: This presents an interesting quagmire: Utilities want to know this information in great detail before they specify a vendor, and vendors are reluctant to reveal any issues that would prevent sales. There is a real urgent need for an information clearing house to capture and publish this information.
Peak: That is why many utilities are seriously considering and developing their own test laboratories. The effort to have a stable, repeatable environment is necessary to establish a baseline of performance. A system-wide, end-to-end test facility will uncover more areas for optimization than each of the piece parts alone. This is notwithstanding the need to test interoperability and perform regression tests on new releases before they are put in the field.
McDonald: So in other words, the meter is only a bit player in the overall solution. Expertise and full testing is needed at a system level.
Dorow: Comparing AMI to SCADA or energy management system (EMS) is an interesting exercise. Today, elements of SCADA and EMS solutions can be tested in isolation. The products offered and the needs they fulfill are mature—they adhere to industry-established and -adopted standards. In comparison, there are relatively few devices in these environments versus in an AMI deployment. In this case, a one percent failure rate is unacceptable and intolerable, but a 0.1 percent failure rate within a deployment of a million meters is significant and can lead to disaster.
As a result, one of the most important process changes that a utility must undertake is rethinking its entire quality assurance program. Standard statistical sampling techniques may be appropriate for mature products; new products need to have additional rigors applied.
McDonald: This brings me back to the issue discussed earlier: One of the lessons learned is that even if the products perform to a uniform 100 percent standard, there is still the physical need to swap out existing meters with the new devices. Unless there is a rigor applied to this process, good accurate meters can be put in service with bad programming.
Bialy: This is particularly true when installations are outsourced to a third-party. Its business driver is the number of installations completed per day. There is little time or experience to build margins into these schedules that would allow on-site customization or any variance from the design standard.
Tajian: That is where the ability to access all the information contained in the meter becomes important. It can be used to validate the configuration and set-up, and in some cases, could even correct any anomalies remotely.
Dorow: With mass deployments, you risk the chance of massive failures, if the necessary steps are not taken before, during, and after the implementation. One slip-up can undo years of preparation.
McDonald: AMI also creates a situation where all the meter-to-bill processes must be reviewed and possibly re-tooled. Automating bad processes just makes the problem that much worse. Utilities must do pre-audits of what needs to be done and make sure that they plan for the unexpected. Technology drives changes in process, and processes must exploit the technology.
Dorow: AMI is business transformational. In the past, cell phones were thought of as a means to communicate away from a fixed location. With the advent of the smartphone, our world has been turned upside-down. The old rules of business logic no longer apply to billing. AMI is a game changer, and utilities must be ready to fully play.
Summary
The complexity of installing smart meters now requires a holistic view of the entire meter-to-bill and ongoing support processes. Older processes will no longer be appropriate—the amount of information that will be available to consumers with AMI puts a new onus on the utility and its customer systems. The informed consumer may now have newer tools at his/her disposal.
With a growing knowledge base of rates, consumption data, and comparative analysis of similar users, progressive utilities must now focus more efforts on the utility-customer relationship. The consumer tolerance for errors and estimates is waning; “good enough” is no longer good enough.
While the challenges of smart grid continue to intrigue power system engineers and designers, the fundamental customer touch point for all of the smartness of the grid is still rooted in the metering function. Utilities need to get this right before customer confidence is lost forever.
About Automation Insight
Automation Insight is a complimentary monthly publication designed specifically for the utility industry and those serving the utility industry. For comments or suggestions on future article topics, please e-mail automation.insight@kema.com.
Subscribe to Automation Insight.